[Note: This page is still under construction and should be used only as a reference. Do not cite without permission.]
World-Magnates Database: Concepts, Measures, and Data Collection
Version 1 – January 2019
Concepts and Measures: Epicenters of Wealth and World-Magnates
This project has involved the identification of and information collection on the richest individuals in the world—those who operated primarily in Fernand Braudel’s “top layer” of the world-economy. These “world-magnates” serve as an indicator of what we call “epicenters of wealth.” Three explanations are required. First, what do we mean by “epicenters of wealth”? Second, what do we mean by “world-magnates”? And finally, how do world-magnates serve as a measure of these epicenters?
“Epicenters of wealth” are hubs of profit creation and accumulation. We discuss epicenters of wealth as embodiments of both spatial-temporal locations and specific wealth generating activities.
In the most obvious way, uncovering world-magnates offers insight because these individuals are the primary beneficiaries of the creation of new epicenters of wealth accumulation. In the Schumpeterian sense, world-magnates represent the cutting edge of creation, providing “early warning of new fortunes into the evolution of industries and markets” (Potts 2006: 348). In the world-systems sense, the economic activities that are generating the wealth of world-magnates can serve as an operationalization of core-like activities.
Creation and Destruction – The world’s economic elite can be considered as an indicator of the spatial and temporal location of epicenters of wealth accumulation. From this point of view, “rich lists are interesting not because of the power structures they ostensibly represent — for they are transitory, disequilibrium phenomena — but rather for the forward insight they give through the early warning of new fortunes into the evolution of industries and markets” (Potts 2006: 348). Thus, these world-magnates serve to map out the unfolding of processes of Creative Destruction. Mapping the world’s billionaires offers insight into processes of Creative Destruction because these individuals are the primary beneficiaries of the creation of new epicenters of wealth accumulation. But the rise and fall of such individuals also helps identify how processes of Creative Destruction are clustered temporally, spatially (e.g., in particular cities, nations or regions) and/or in specific production, trade and investment networks.
Core-like Activities – Several recent studies of nation-state position in the world-system—whether in hierarchies of wealth or networks of trade—have refocused attention on decades-old questions regarding the criteria used to classify territories into the categories of “core,” “semiperiphery,” and “periphery” (Karatasli 2017; Karatasli and Kumral 2018; Pasciuti and Payne 2018; Clark and Beckfield 2009; Mahutga 2006; Jorgenson 2006; Babones 2005; Kentor 2000). Undergirding such discussions are serious debates about what is meant by these terms and the assumptions made in applying them—particularly in historical work. Wallerstein argued that core regions of the world-economy specialize in high-profit, high-wage economic activities (core-like), which derive the most benefits from the world division of labor. Peripheral regions, by contrast, specialize in low-profit, low-wage activities (peripheral), deriving the least benefits from the world division of labor. Between them, semiperipheral areas are characterized by a more or less even mix of core and peripheral activities.
Arrighi and Drangel (1986) attempted to operationalize “core” and “peripheral” not by classifying the activities themselves, but rather by identifying the economic consequences of those activities. Because core-like activities yield greater profits than peripheral activities, “residents of the [core] must command a large share of the total surplus produced in the world-economy while residents of the [periphery] must command little or no such surplus… this difference must be reflected in a…differential between the per capital GNP of residents in the two types of states” (Arrighi 1985: 244). Arrighi and Drangel (1986) found that this was, in fact, the case: for the mid-twentieth century, nation-states clustered into three income zones, corresponding to the core, periphery, and semiperiphery. This finding was replicated and extended (Korzeniewicz and Martin 1994; Arrighi, et. al. 1996; Babones 2005), lending credence to the idea that such arrangements were a permanent feature of the world-economy (Wallerstein 2005: 1267). Karatasli (2017), using a similar method, has found that this structure is far more subject to change than originally assumed. He has identified transformations in the world-wealth hierarchy that correspond with crises of the capitalist system—crises that yield a transformation in the forces of production and a subsequent rearrangement of not only core and peripheral locales, but also core and peripheral activities (see also: Silver 2003; Karatasli and Kumral 2018). These transformations, explored by Karatasli (2017) and implicitly theorized in the operationalization by Arrighi and Drangel (1986), guide our work: The activities that are core-like are constantly transforming and rearranging in a process of Schumpeterian “creative destruction.”
We can uncover these core-like activities and spaces—and their transformations—by identifying the most profitable economic activities in the world-economy at any point in time. Using our data, we argue, allows for a less problematic review of these activities by moving the unit of observation away from the nation-state and to world-magnates—a shift which better allows for the maintenance of the world-system as unit of analysis. By tracing the rise and fall of the wealth-generating economic activities of these individuals, we are able to identify how the most profitable enterprises are clustered temporally, spatially and in specific production, trade, and investment networks. Such an operationalization builds from Arrighi (1985), Arrighi and Drangel (1986), and Karatasli (2017) in that it utilizes the economic impacts of the core-like activity to map world-systemic processes. Unlike these studies, however, this operationalization utilizes the economic impacts of core-like activities to identify the activities themselves, as opposed to the territorial boundaries in which they are contained.
Proceeding in this way, our data allow us to overcome a priori assumptions about time, space, and industry. Rather than assuming either the spatial-temporal location or the type of activities involved in the most profitable economic epicenters, our data allow us to proceed in the opposite direction, using the data on world-magnates to identify major shifts in the epicenters of wealth accumulation—in both the specific activities involved, and in their spatial and temporal clustering.
Once potential entries have been identified (see section III below), relevant information on the world-magnate is identified. Using secondary sources focusing on the case at hand, and primary sources as available and necessary, our research collects individual information on characteristics such as year of birth, birth of death, place of birth, place of residence, place of operation, place at death, the estimated extent of wealth by measures at the time, the type of activities that generated this wealth, the extent of connection and/or involvement with local and national governments, the extent to which inheritance was a source of the individual’s wealth. In addition, we collect information on the linkages of each entry to other individuals in our database (through kinship or otherwise). We have found that the collection of these data is an iterative process, as research on other individuals often produces new information that can be added to existing entries.
The collected data provide the necessary information to construct 17 variables in three main variable groups in the final database:
Level of Wealth (absolute billionaire, abs billionaire flag, relative billionaire, relative billionaire flag). The most important, and challenging characteristic to codify is individual net worth. Ideally, and in many cases thus far, we can find a reliable estimate of the individual’s peak wealth. Because wealth is typically reported in the local and contemporary currency, we adjust reported wealth for inflation and convert to a standardized currency using the exchange rate.
We use two distinct metrics to identify the historico-structural equivalents of today’s “billionaires.” The first, absolute wealth, identifies billionaires as having a net worth that exceeds $1,000,000,000 in 2000 US$. The second, relative wealth, compares the ratio of individual net worth to contemporary global GDP against the same ratio for a dollar billionaire in the year 2000. We frequently collect information on a wealthy individual that falls below both thresholds; she is not a billionaire in terms of her absolute wealth or in terms of her wealth relative to the size of the global economy at the time. We retain those names in the survey, as that information may be useful later, but data collection and verification efforts will focus on those individuals that cross at least one billion threshold.
In some cases, our data collection process identifies a promising individual, but is difficult to determine a reliable measure of individual wealth. These cases usually fall in one of two categories. In the first we are presented with a long list of assets (land ownership, ships, art collection, etc.). In the second, wealth is expressed in relative terms (e.g., among the most wealthy men in Amsterdam or Mexico). In each case, we aim to generate a quantitative estimate of wealth for analysis. To this end, we identify in each case a comparable or set of comparables, individuals with a similar claim on assets or also among the “wealthiest men of Amsterdam or Mexico,” for whom there are reliable and quantitative estimates of individual wealth. Our database flags cases for which individual net worth is estimated using this technique. In addition to the dichotomous measures of billionaire status, we include secondary variables to indicate a confidence level that the individual is, indeed, a billionaire.
Demographics (birth year, birth country, death year, sex, religion, inherited). Inheritance, in turn, is comprised of four categories: wealth is inherited and relatively dormant; wealth is inherited and active; some wealth is inherited but principal added significantly to that wealth; inheritance is insignificant.
Field of Accumulation (industry 1, industry 2, operation c, operation 1, operation 2, relationship to state). We recognize 21 industries or sectors. Each individual is coded with a principal industry (industry 1), and we identify a second industry if it is appropriate (industry 2). We also locate a principal country of operation (contemporary [operation c] and modern [operation 1]) and list other areas of operation (operation 2) as a string variable. Relationship to the state recognizes six potential identities: aristocracy, state employee, unique economic privilege (e.g., state monopoly), major trade partner, substantial financial dealings, no immediate relationship to the state.
In short, the level of wealth variables are essential to evaluate whether or not a case should be included and compare the extent of wealth managed from top wealth holders over time; the demographic variables allow us to measure individual characteristics of world-magnates; the field of accumulation variables are indicators of sites and strategies of wealth accumulation. In addition to the codified results for statistical analysis, we will include the relevant details as notes (e.g., a list of the individual’s assets) and relevant source information. Simultaneously, we build as well a network array of relationships (family ties, business ties) between magnates.
This project has become feasible only recently. Because the individuals included in the survey are immensely wealthy, in many cases their identities and basic demographic information have left a record that is often accessible through appropriate bibliographic searches. However, a project such as this would have required an extraordinary number of research hours even a decade ago, as collecting the relevant data requires detailed surveys of a vast field of secondary and primary literatures in multiple languages. New technologies, however, have simplified some of the tasks at hand. Research engines now allow researchers to individually review extensive bibliographies (in English and other languages) in a shorter timeframe. Moreover, the recent availability of translation engines online expands the access of every individual researcher to a bibliography in a large number of languages.
A team of trained researchers has been tasked with identifying world-magnates. We have used several procedures to initially identify these individuals:
- Our research team draws many names from the extensive bibliography available on the general subject of the very wealthy. In some cases, the existing literature provides lists of the very wealthy for particular periods of time and/or for particular geographical locations (for example, Lundberg (1968) provides extensive lists of the very rich in the United States in the immediate post-WWII era; Beresford and Rubinstein (2011) have been providing historical lists for Great Britain);
- Our research team also has drawn many additional potential entries by conducting extensive bibliographical research entering key words (e.g., rich, wealthy, magnates, millionaires, billionaires) and both broad and particular areas of the areas of the world (e.g., Latin America, Asia, China, Poland, London, Bombay, Mumbai) in worldwide library catalogues, biographical encyclopedias (such as various versions of Who’s Who, the Dizionario Biografico Degli Italiani and the 1906 Jewish Encyclopedia) and search engines (such as Google’s Ngram). This strategy has proven to be particularly useful for uncovering potential entries in areas other than Western Europe or the United States;
- We review the literature on key production, trade and consumption networks (e.g., the whale oil industry in the early nineteenth century; shipping; agricultural commodities; mineral oil production and refining; automobile production; finance), paying particular attention to areas of the world that might be underrepresented in the existing literature on the very wealthy, to identify key individual actors within these networks. This strategy also has been particularly useful for uncovering potential entries in areas other than Western Europe or the United States;
- All these potential entries themselves serve as the basis for extending a version of “snowball sampling”: as we conduct research on individual entries, we often uncover a network of additional individual who become potential entries into our dataset. For example, research on the financial networks of the Rothschilds in the nineteenth century (using sources such as Ferguson 1998 and 1999) can identify potential entries such as Henry Hope (1735-1811) of Amsterdam or William IX of Hesse-Kassel (1787-1867). Moreover, this is an iterative process, as each new entry potentially provides insights into new networks (and when they cease to do so, when their networks involve other individuals who are already in our database, we receive an indication that our search has been fairly exhaustive).
Reliability of Data
In assessing our reliability, we understand that the coverage we can expect in a historical and global dataset of world-magnates will always be incomplete (in the same way that existing “maps” of the universe are always incomplete). But we do want to assess whether the coverage of our dataset is as inclusive as possible of the available information (again, knowing as well that new information is likely to emerge in the future as experts in specific times and places evaluate our coverage). We use several procedures to test the coverage of our unique data:
- We continuously track the number of names generated by our searches that are already included in our database. This procedure offers an opportunity to track over time the extent to which our coverage becomes more exhaustive;
- We closely examine the temporal and spatial patterns revealed by our database. We have learned from our data collection that turbulent fluctuations in the number of very wealthy individuals within any given location in a short number of years serve to highlight points in time and place that require greater research to ascertain whether these fluctuations are the product of inadequate data or actual transformations in our field of study. Examining these patterns on a regular basis help us direct our data collection efforts more effectively;
- We regularly conduct a self-test. At regular intervals, researchers dedicate a fixed number of hours towards identifying very wealthy individuals without reference to the existing database. Generally, they will be assigned specific geographic regions and time periods (as drawn from the study of patterns mentioned above) to focus their search. When they have completed their search, we compare those results against the existing survey to produce coverage estimates by period, region, and “level of wealth” (i.e., the relative wealth of the individual). This self-test allows us to assess the extent of coverage ensured by our data collection methods, and helps reveal the potential temporal/spatial biases of our data collection methods.